Hierarchical Prototype Network for Interpretable Chest X-ray Disease Classification

References

[1] K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778, 2016.

[2] C. Chen, O. Li, D. Tao, A. Barnett, C. Su, and C. Rudin, “This Looks Like That: Deep Learning for Interpretable Image Recognition,” Advances in Neural Information Processing Systems (NeurIPS), vol. 32, pp. 8930–8941, 2019.

[3] O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation,” Medical Image Computing and Computer-Assisted Intervention (MICCAI), pp. 234–241, 2015.

[4] X. Wang, Y. Peng, L. Lu, Z. Lu, M. Bagheri, and R. Summers, “ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases,” IEEE CVPR, pp. 2097–2106, 2017.

[5] A. Rajpurkar et al., “CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning,” arXiv preprint arXiv:1711.05225, 2017.

[6] G. Litjens et al., “A Survey on Deep Learning in Medical Image Analysis,” Medical Image Analysis, vol. 42, pp. 60–88, 2017.

[7] R. Tjoa and C. Guan, “A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI,” IEEE Transactions on Neural Networks and Learning Systems, vol. 32, no. 11, pp. 4793–4813, 2021.

[8] S. Bach et al., “On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation,” PLOS ONE, vol. 10, no. 7, 2015.

[9] R. R. Selvaraju et al., “Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization,” IEEE International Conference on Computer Vision (ICCV), pp. 618–626, 2017.

[10] Y. Zhou et al., “Prototype-Based Interpretable Deep Learning for Medical Image Diagnosis,” IEEE Journal of Biomedical and Health Informatics, vol. 25, no. 9, pp. 3423–3433, 2021.

[11] S. Lundberg and S. Lee, “A Unified Approach to Interpreting Model Predictions,” Advances in Neural Information Processing Systems (NeurIPS), pp. 4765–4774, 2017.

[12] J. Shin et al., “Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning,” IEEE Transactions on Medical Imaging, vol. 35, no. 5, pp. 1285–1298, 2016.

[13] P. Mooney, Chest X-ray Pneumonia Dataset, Kaggle, 2018. [Online]. Available: https://www.kaggle.com/datasets/paultimothymooney/chest-xray-pneumonia

[14] A. Esteva et al., “Dermatologist-level Classification of Skin Cancer with Deep Neural Networks,” Nature, vol. 542, pp. 115–118, 2017.

[15] T. A. Al-Antari et al., “Fast Deep Learning Computer-Aided Diagnosis of COVID-19 Based on Digital Chest X-ray Images,” Applied Intelligence, vol. 51, pp. 2890–2907, 2021.

[16] J. G. Silva et al., “Explainable Deep Learning for Chest X-ray Image Classification,” IEEE Access, vol. 8, pp. 180034–180045, 2020.

[17] A. Esteva et al., “Dermatologist-level classification of skin cancer with deep neural networks,” Nature, vol. 542, no. 7639, pp. 115–118, 2017.

[18] G. Litjens et al., “A survey on deep learning in medical image analysis,” Medical Image Analysis, vol. 42, pp. 60–88, 2017.

[19] H. R. Roth et al., “Deep learning and its application to medical image segmentation,” Medical Image Analysis, vol. 45, pp. 1–19, 2018.

[20] J. Irvin et al., “CheXpert: A large chest radiograph dataset with uncertainty labels and expert comparison,” AAAI Conference on Artificial Intelligence, 2019.

[21] M. T. Ribeiro, S. Singh, and C. Guestrin, “Why should I trust you? Explaining the predictions of any classifier,” KDD, 2016.

[22] S. M. Lundberg and S. I. Lee, “A unified approach to interpreting model predictions,” NeurIPS, 2017.

[23] B. Kim, M. Wattenberg, J. Gilmer, et al., “Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (TCAV),” ICML, 2018.

[24] Y. Zhou, Y. Li, and W. Li, “Interpretable deep learning models for medical image diagnosis,” IEEE Journal of Biomedical and Health Informatics, 2021.

[25] [D. G. T. Denil et al., “Learning where to attend with deep architectures for image tracking,” Neural Computation, 2012.

@article{1722241,
      author = {Akshaya Sreekumar, Sankalp Kumar Singh, Dr. Sibi Amaran, R. Shobana},
      title = {Hierarchical Prototype Network for  Interpretable Chest X-ray Disease Classification},
      journal = {Iconic Research And Engineering Journals},
      year = {2026},
      volume = {10},
      number = {2},
      pages = {941-949},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1722241.pdf},
      abstract = {In recent years, deep learning has made great strides in medical image categorization, especially for automated chest X-ray image analysis for illness screening and early detection. Even though they are quite good at producing predictions, most deep learning-based models behave as black-box decision-making systems that don't give much information about how an input is classified. This could make clinician trust and interpretability weaker. To address this challenge, this research presents a Hierarchical Prototype network (HierProtoPnet) for interpretable chest X-ray classification of various images which are provided.nThe suggested solution uses a ResNet50 backbone for strong feature learning and two levels of prototype learning. This lets it learn both fine-grained and coarse-grained visual patterns in medical images at the same time. Lower-level prototypes learn more detailed local properties, such textures and edge patterns. Higher-level prototypes, on the other hand, encode more abstract disease aspects. Because of its hierarchical structure, the model classifies based on learnt prototypes, which gives clear predictions.
The suggested method uses a ResNet50 backbone for strong feature extraction and has two levels of prototype learning. This allows it to learn both little and large visual patterns in medical images at the same time. Prototypes at a lower level learn more about the finer points of a place, like textures and edge patterns. Higher-level prototypes represent broader and more general disease characteristics. Because of this hierarchical structure, the model makes predictions by comparing input images with learned prototypes. This makes the reasoning behind each prediction clearer and easier to understand. The prototype-based approach also improves interpretability by showing how different visual patterns influence the final decision. The results indicate that hierarchical prototype learning supports diagnosis while improving both model accuracy and clarity, strengthening its suitability for clinical decision support systems. This study highlights the potential of combining deep convolutional feature extraction with hierarchical interpretability mechanisms to create accurate and informative models for medical image classification.},
      keywords = {Chest X-ray, Deep Learning, Hierarchical Prototype Network, ProtoPNet, ResNet50, Medical Image Classification, Explainable Artificial Intelligence (XAI)},
      month = {August},
      doi = {https://doi.org/10.64388/IREV10I2-1722241}
  }